Re: glibc 2.0.103

[Thorsten Kukuk <kukuk@weber-eb.uni-paderborn.de>]

> 
> >>>>> Ulrich Drepper writes:
>  >[...]
>  > Major changes since last week:
>  > - general cleanup
>  > - incorporate latest PPC changes (no reported error in the testsuite anymore!)
>  > - better patch for RPC DOS
>  > - glibc should now be select()-free
> 
> The last two(?) missing pieces are:
> sunrpc/svc_run.c:      switch (__select (_rpc_dtablesize (), &readfds, (fd_set *)NULL
> nis/nis_callback.c:      switch (select (_rpc_dtablesize (), &readfds, NULL, NULL, &tv))
> 
> if we like to get rid of these, we need to convert the contents of
> svc_fdset.

The problem is, that a lot of daemons overwrite svc_run (). The only solution
I see is, to implement a second version of this svc_* functions, which uses
svc_fdset. The new functions and the new svc_run () will use poll, old
daemons should work fine with the old svc_* versions. I think FreeBSD
had make something similar with the dynamic svc_fdset patches.

nis_callback.c will be a problem, I don't know what happens, if we use
a select base svc_run with a poll based nis_callback. Need to check it.

  Thorsten
-- 
Thorsten Kukuk  kukuk@vt.uni-paderborn.de
                http://www-vt.uni-paderborn.de/~kukuk/
Linux is like a Vorlon.  It is incredibly powerful, gives terse,
cryptic answers and has a lot of things going on in the background.